# On Causes, Diagrams, and the Arrogance of the Explicable Intelligence is the faculty of knowing what matters and *why*. Not the mere cataloguing of correlations—any clerk with a ledger can do that—but the piercing through of surfaces to the actual mechanisms by which one thing compels another. To lack this is to be forever at the mercy of appearances, which is to say, to be a fool. Yet our so-called scientific establishment spent a century—a full century!—expelling causation from its vocabulary as though it were a contaminating spirit. The excuse was rigor. The result was intellectual cowardice dressed in mathematical formality. Correlations, they said, are all we can know. Causation is metaphysics. Better to hide in the safe harbor of *p*-values than to risk claiming that one thing *makes* another happen. This was nonsense, and Judea Pearl had the intellectual courage to say so plainly. Pearl's contribution is real and muscular: he gave us the *do*-calculus, the directed acyclic graph, a formal apparatus for reasoning about interventions rather than mere observations. When you manipulate a variable, you learn something fundamentally different from what nature's uncontrolled patterns can teach. This is true. His diagrams make explicit what was previously left to intuition and guesswork. For that, he deserves credit. But—and here is where the modern scientist's self-deception begins anew—**explicitness is not truth**. A diagram is a *claim*, not a discovery. It is an act of interpretation, an argument made in the language of arrows and nodes. And interpretation is where all the blood-and-bone problems of knowledge actually live. You can draw your causal diagram with absolute clarity, with every assumption labeled and every variable named, and still be *completely, catastrophically wrong*. Clarity is the enemy of humility. This matters most acutely—indeed, it matters *only*—when we venture into the social. --- ## The Diagram as Tyranny When Pearl's apparatus is applied to physical systems—to genetics, to pharmaceutical interventions, to mechanical processes—the ground beneath the diagram has some solidity. Nature has already decided the causation. The diagram merely represents an external fact. When you ask "Does this drug cause this outcome?" you can, at least in principle, run an experiment. You can interfere. You can *do*. Nature will answer. But when we move into the social—into intelligence, education, poverty, crime, the human texture of ordinary life—the diagram becomes something far more dangerous: it becomes *legislation*. Who decides what causes intelligence? The researcher who draws the diagram. Who decides whether intelligence is caused by genes, or schooling, or nutrition, or the social position of one's parents? The diagram-maker. Who decides which arrows point where? The person holding the pen. And that person brings to the task every prejudice, every unstated assumption, every invisible inheritance of their own social position. Consider a simple example: a diagram linking parental income to children's test scores, with genes as a mediating variable. It looks innocent enough. The arrows are drawn. The math follows. A researcher can be completely explicit about the mechanism, can publish it, defend it, cite it a thousand times. And the diagram can be *completely wrong*—not because the math is faulty, but because the actual causal structure of human life is far more intricate and historically contingent than any diagram can capture. The problem is not that we made the diagram; it is that we believe the diagram speaks for nature rather than for ourselves. Genes do not float free from society. They express themselves within particular material conditions, particular histories of advantage and disadvantage, particular structures of power that have been centuries in the making. To draw an arrow from genes to test scores *as though these were independent variables* is to smuggle in a lie. It is to treat what is socially constructed as though it were natural. It is to mistake your diagram for the world. And once you have done that—once you have made your assumptions explicit and mathematical—you have actually made it *harder* to see the lie. The explicitness becomes a kind of armor. Critics must now attack not your reasoning (which may be flawless) but your premises. And when they do, you can always retreat behind the claim that you stated your assumptions plainly. "I was transparent," you say. "If you disagree, propose a different diagram." But this misses the essential point: **the choice of what to diagram is not a scientific choice. It is a choice about what the world is.* --- ## Knowledge and Power To know the cause of something is not merely to describe a mechanism. It is to hold power over it. If intelligence is *caused* by genes, then society's job is to sort people according to their genetic endowment. If intelligence is caused by schooling, then society's obligation is to equalize schools. If it is caused by the material conditions of early childhood, then the whole edifice of social priority shifts. The diagram is not innocent. It is a blueprint for action. And who gets to draw it? In practice, it is drawn by those who can afford the equipment, who work within institutions, who publish in the right journals, whose assumptions have already been validated by the existing order. The person asking "Are intelligence differences caused by genes or environment?" already inhabits a world structured by answers to that very question. Their lived position shapes what they can see. Their diagram will reflect it. This is not an argument against causal reasoning. It is an argument against the *pretense* that causal reasoning, once made explicit and mathematical, is thereby purified of perspective. The worst thing Pearl's apparatus has done—and I say this with genuine respect for his technical accomplishment—is to make the imposition of perspective look like the discovery of truth. The researcher can now say: "Here is my diagram. Here are my assumptions. The math is sound. Therefore my conclusions are valid." And the person who asks, "But who decided those assumptions? What would a different diagram show? What is being hidden by choosing to diagram things this way?"—that person appears to be attacking science itself, rather than attacking a particular exercise of power. --- ## The Social Dimension: Where Diagrams Fail In the social realm, causation is *constituted* rather than discovered. Poverty causes poor health outcomes—but poverty is not a natural fact; it is a set of legal, economic, and social arrangements that could be otherwise. Intelligence is correlated with educational opportunity—but educational opportunity is not distributed by nature; it is distributed by human decision. To ask "What causes intelligence?" without first asking "What do we mean by intelligence? Who benefits from measuring it this way? What is being ignored?" is to begin with an answer already built into the question. Yet the diagram-maker does precisely this. The diagram *assumes* that intelligence is a coherent, measurable thing. It assumes that test scores capture something real about human capacity. It assumes that these scores are caused by identifiable factors that can be isolated and studied. All of this is contestable—not as a matter of opinion, but as a matter of historical and social fact. Intelligence tests were developed by particular people in particular moments to serve particular purposes. Their history is entangled with racism, with the desire to sort and manage populations, with the ambition to naturalize inequality. A diagram that takes intelligence as its starting point has already made a political choice. It has chosen to accept the categories of the powerful as the categories of nature. This does not mean we should abandon causal reasoning in the social domain. It means we should be ruthlessly honest about what we are doing when we draw a diagram. We are *proposing* a structure. We are *arguing* for a particular way of seeing. We are *deciding*—often without acknowledging it—what counts as a cause and what counts as an effect, what is a variable and what is a background condition, what is susceptible to intervention and what is fixed. And we should be prepared to defend that choice not by appeal to mathematical rigor, but by appeal to its consequences: What world does this diagram help us build? Who benefits from seeing causation this way? What alternatives are being foreclosed? --- ## The Verdict Intelligence, properly understood, is the capacity to see through the diagram to the interests it serves. It is the refusal to mistake explicit reasoning for unbiased reasoning. It is the ability to hold in mind, simultaneously, that Pearl's tools are genuinely powerful *and* that they can be used to dress up prejudice in the language of necessity. The researcher who draws a diagram without acknowledging the violence implicit in that act of drawing—who presents assumptions as if they were facts, who claims transparency when what they have achieved is only explicitness—that researcher is not intelligent. He is merely fluent. And fluency in the language of science is among the most effective ways to prevent intelligence from doing its actual work. To know a cause is to know it provisionally, contextually, in full awareness of the perspective from which you are knowing it. Anything less is not rigor. It is just another form of blindness, dressed up in mathematics.